KR101698033B1 - Apparatus and Method for correcting Cone Beam Artifact on CT images - Google Patents

Abstract

The present invention relates to a cone beam artifact correction apparatus and method capable of extracting information on a high density material (bone), which is a main cause of cone beam artifacts using dual energy, and improving the performance of a cone beam artifact correction algorithm that has been proposed by using an iterative reconstruction technique , Which is a dual energy technology analysis that separates the projection data obtained by using two different energies into projection data of only each constituent material including the Bone SaNogram and Water Cygnogram through the material decomposition algorithm The bone reconstruction image is estimated through image restoration using a mathematical iterative reconstruction method with the bone segmentogram, and transformed into projection data (Raw Data) obtained through CT imaging with the estimated bone reconstruction image After the Forward Projection process, Acquiring a reconstructed FDK bone reconstructed image of the dense material exhibiting the reconstructed FDK bone artifact, acquiring the reconstructed cone beam artifact reconstructed image through a difference between the Bone reconstructed image and the FDK bone reconstructed image, And a cone beam artifact correction unit that finally obtains a cone beam artifact correction image through a combination.

Description

[0001] Apparatus and Method for Correcting Cone Beam Artifacts on CT Images [

The present invention relates to an apparatus and method for correcting cone beam artifacts in a CT image, and more particularly, to a cone beam artifact correction method and apparatus for correcting cone beam artifacts in a CT image using a large area X- And more particularly,

Cone beam artifacts are a well known problem in circular orbital computer tomography. Toshiba's CT equipment, AquilionONE, etc., has a very large cone angle of the X-ray source of the high CT system and the data of the radon area is defective. Therefore, the cone beam is used as a CFK (Carbon Fiber Composite) Artifacts are created.

As a method for correcting the current cone beam artifact, there are a first method of algorithmically correcting the deficit data and a second method of changing the data acquisition method of the cone beam CT.

And FDK (Feldkamp Davis Kress) algorithm, which is widely used for image reconstruction of cone beam CT, increases cone beam artifacts as the angle of cone beam increases. In order to compensate for defective data, there are a P-FDK (Parallel-FDK) and a T-FDK (Tent-FDK) method for transforming cone beams into parallel beams for image reconstruction, There is a two pass algorithm that reproduces and corrects from the image.

That is, when the projection structure of the cone beam CT is given as shown in Fig. 1 (a), the P-FDK converts the 3D cone beam data into 3D parallel beam data as shown in Fig. 1 (b) And reconstructs the image. Also, as shown in FIGS. 1 (c) and (d), T-FDK reconstructs a tent-like projection structure based on P-FDK to generate a virtual detector.

2, a two-pass algorithm is a method in which a high-density material (bone), which is a main cause of artifacts in the cone beam artifact image 10 generated by using the FDK algorithm in an existing cone beam CT system, We estimate the bone segmentation image (20) by segmentation process by numerical comparison in image. This means that the estimation performance of the high-density material is variable depending on the arbitrary threshold value.

A forward projection process for converting the Bone segmented image 20 of the high density material estimated from the CT image into projection data (Raw Data) obtained through CT imaging, and a forward projection process for reconstructing the image using the FDK algorithm, Obtain an FDK bone reconstruction image (30) in which the artifact appears.

Then, the reproduced cone beam artifact reconstructed image 40 is obtained through the difference between the Bone segmented image 20 and the FDK bone reconstructed image 30.

Thereafter, the cone-beam artifact-corrected image 60 is finally obtained through a combination of the cone-beam artifact-generated image 50 and the cone-beam artifact reproduced image 40.

Cone beam artifacts are mainly focused on high density materials. That is, a high-density material is separated through a threshold method through numerical comparison in an image in which a cone beam artifact appears, and additional projection and image reconstruction are performed on the material to reproduce cone beam artifacts due to a high-density material. The reproduced cone beam artifact is removed from the original image to correct the cone beam artifact.

This conventional algorithm-based cone beam artifact correction method is applied to a CT system (e.g., a 320 slice Toshiba CT system, a cone beam CT system for radiotherapy monitoring, and the like) having a cone beam angle of 10 degrees or more at a small cone angle , C-ARM CT system), artifact correction effect is reduced, and there is a limit to image quality improvement.

Conventional cone beam CT images the object by rotating the X-ray source and detector along a center circle. In order to reduce cone beam artifacts through acquisition of defect data, methods of obtaining data on objects by changing the shooting trajectories of source and detector have been studied. As shown in FIGS. 3A and 3B, the source and the detector move according to the circle-and-line and the saddle orbit, and the 3D image obtained by reducing the cone beam artifacts Restoration is possible.

As another method, a system structure using several X-ray sources such as MS-IGCT (X-ray Source-Inverse Geometry CT) as shown in (c) . This makes it possible to obtain sufficient projection data by supplementing the defect data generated by one X-ray source and detector with an additional X-ray source.

However, the method of removing cone beam artifacts by changing the data acquisition method is theoretically possible, but it is difficult to apply to the currently used CT system because the device structure itself needs to be changed. Further, the photographing time is increased due to the change of the photographing trajectory, so that there arises a problem that additional artifacts occur due to the movement of the patient.

Japanese Patent Application No. 2011-141209 (filed on June 24, 2011) Japanese Patent Application No. 2012-232359 (filed on October 19, 2012)

SUMMARY OF THE INVENTION Accordingly, the present invention has been made to solve the above-mentioned problems, and it is an object of the present invention to provide a method and apparatus for extracting information on a high density material (bone), which is a main cause of cone beam artifacts, Which can improve the performance of the cone beam artifact.

Other objects of the present invention are not limited to the above-mentioned objects, and other objects not mentioned can be clearly understood by those skilled in the art from the following description.

According to another aspect of the present invention, there is provided an apparatus for correcting cone beam artifacts in a CT image, comprising: a projection unit configured to project projection data obtained by using two different energies into a bone- A Bone reconstruction image is estimated through image restoration using a mathematical iterative reconstruction method using the Bone interogram, and the estimated Bone reconstruction image A forward projection process for transforming the data into raw data that can be obtained through a CT scan and an FDK algorithm for reconstructing the FDK bone reconstructed image to obtain an FDK bone reconstructed image having a cone beam artifact, Reconstructed Cone Beam Artifacts through Differences from FDK Bone Reconstructed Images Obtaining a phase, and may consists, finally includes a cone beam artifact correction to obtain the cone beam artifact corrected image through a combination of the cone beam artifact generated images and cone beam artifacts reproduced image.

Preferably, the two different energies are characterized by having two energy spectra including a low energy spectrum and a high energy spectrum.

Preferably, the material disassembly algorithm constructs a polynomial through the projection data Y1 and Y2 already known through the dual energy spectrum to obtain a value of the basis function to be obtained, And estimation is performed by a polynomial coefficient estimation and a least square fitting process through a calibration phantom.

Preferably, the mathematical iterative reconstruction method uses the sineogram value matrix b and the matrix A related to the geometry of the system used to obtain the data to calculate the total variation : TV) is minimized through an optimization process and an iterative process.

According to another aspect of the present invention, there is provided a method of correcting cone beam artifacts in a CT image, the method comprising: (A) analyzing projection data obtained using two different energies through a dual energy technology analysis unit, (B) separating the projection data of the separated high-density material (bone) through the cone beam artifact correction unit into projection data of the respective constituent materials including the sinogram and the water- Estimating a Bone reconstruction image through image reconstruction using a mathematical iterative reconstruction method; and (C) reproducing and correcting the cone beam artifact based on the Bone reconstruction image of the high density material estimated through the iterative reconstruction image reconstruction .

Preferably, the two different energies are characterized by having two energy spectra including a low energy spectrum and a high energy spectrum.

Preferably, the material decomposition algorithm includes a step of constructing a polynomial through the projection data Y1 and Y2 already known through the dual energy spectrum, Estimating a coefficient of a polynomial through a calibration phantom and a least square fitting process.

Preferably, the mathematical iterative reconstruction method uses the sineogram value matrix b and the matrix A related to the geometry of the system used to obtain the data to calculate the total variation : TV) is minimized through an optimization process and an iterative process.

Preferably, the step (C) includes a forward projection process for converting a bone reconstruction image of a high-density material estimated through the restoration of the repeated reconstruction image into projection data (Raw Data) Acquiring an FDK bone reconstruction image in which the Cone beam artifact appears by performing image reconstruction using the FDK algorithm, acquiring a cone beam artifact reconstructed image reproduced through the difference between the Bone reconstruction image and the FDK bone reconstruction image, And obtaining the cone beam artifact corrected image through a combination of the obtained concave artifact reproduced image and the obtained concave artifact reproduced image.

The apparatus and method for correcting cone beam artifacts in a CT image according to the present invention as described above have the following effects.

First, unlike conventional Cone Beam Artifact Correction Technology, it is expected that it will be readily available to the currently used CT system and can be used easily.

Second, it can be applied to a large-area X-ray detector that generates a large cone angle because the accuracy and performance of Cone beam artifact is superior to that of existing Cone beam artifact correction technology.

Third, because of the excellent cone beam artifact correction performance through precise estimation of high density material which is the main cause of cone beam artifact and mathematical iterative reconstruction process for estimating high density material, it is much more superior than the existing proposed algorithm even in a noisy environment It is expected that it will perform.

1 is a schematic diagram illustrating a conventional cone beam artifact correction technique;
FIG. 2 is a block diagram showing a cone beam CT projection structural modification in a conventional Cone beam artifact correction technique
FIG. 3 is a schematic diagram showing a data acquisition method and a system structural change in a conventional Cone beam artifact correction technique;
FIG. 4 is a diagram illustrating a process of a method for correcting cone beam artifacts in a CT image according to an embodiment of the present invention.
5 is a diagram showing a cone beam CT system used in simulation for a cone beam artifact correction technique in a CT image according to an embodiment of the present invention
FIG. 6 is a graph showing the actual deflation phantom damping coefficient distribution used in FIG. 5; FIG.
7 to 10 are views showing a reference image and a cone beam artifact generation image used in a simulation using a conventional Cone beam artifact correction technique
11 is a view showing a reference image and a cone beam artifact generation image used in a simulation using the cone beam artifact correction technique according to the present invention
Fig. 12 is a graph comparing the results of the present invention with the conventional two-pass algorithm

Other objects, features and advantages of the present invention will become apparent from the detailed description of the embodiments with reference to the accompanying drawings.

A preferred embodiment of an apparatus and method for correcting cone beam artifacts in a CT image according to the present invention will be described with reference to the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. It is provided to let you know. Therefore, the embodiments described in the present specification and the configurations shown in the drawings are merely the most preferred embodiments of the present invention and are not intended to represent all of the technical ideas of the present invention. Therefore, various equivalents It should be understood that water and variations may be present.

4 is a block diagram illustrating a method of correcting a cone beam artifact in a CT image according to an embodiment of the present invention.

Referring to FIG. 4, in the present invention, high-density material (bone), which is a main cause of cone beam artifact generation in the conventional two-pass algorithm, is not estimated by numerical comparison in CT image, We extract information about high density material (bone) which is main cause of artifact and improve performance of Cone beam artifact correction algorithm which has been proposed by using iterative reconstruction technique.

In more detail, the dual energy technology analysis unit 100 includes the Bone Cyberogram 110 and the Water Cyberogram 120 through the material decomposition algorithm, the projection data obtained using two different energies Into projection data of only the respective constituent materials.

The material decomposition algorithm will be described as follows.

Basically, the X-ray attenuates when passing through the material, and the degree of attenuation is expressed by the X-ray attenuation coefficient. At this time, there are two physical phenomena that occur when the X-ray passes through the material: photo-electric effect and Compton effect, so that the damping coefficient of all materials is physically converted to photoelectric It can be expressed as a linear combination of two basis functions that correspond to the effect and the compton effect. In the mathematical transformation process, the two basis functions corresponding to the photoelectric effect and the compton effect can be transformed into arbitrary different materials. The material decomposition algorithm is an algorithm for finding the basis functions of two substances . At this time, since the number of basis functions, which is a desired solution, is 2, at least, the values of two or more input X-ray projection data must be known to solve the equation. In this case, Projection data of the material can be obtained through data obtained using an energy spectrum (generally a low energy spectrum, a high energy spectrum). That is, the projection data of a desired substance can be reconstructed by utilizing two measurement data and an X-ray attenuation coefficient of a previously known substance. However, due to the polychromatic nature of X-rays (the property that the X-ray attenuation coefficient of a material continuously changes according to Photon's energy spectrum), estimation of values through direct matrix inversion is error- We construct a polynomial based on the projection data Y1 and Y2 already known through the dual energy spectrum, and use a calibration phantom with a known value of the baseline function. (Polynomial fitting) algorithm, which is a calibration phantom-based polynomial fitting method.

Bone cyberogram 110 among the projection data of the high-density material (bone) separated from the dual energy technology analysis unit 100 through the cone beam artifact correction unit 200 is subjected to image restoration using a mathematical repetitive reconstruction method The reconstructed image 210 is reconstructed as a reconstructed image.

The mathematical repetitive reconstruction method will be described in detail as follows.

In order to reconstruct a bone tomogram obtained through the dual energy technique analysis unit 100 and finally to perform image reconstruction with only a high density material (bone), a mathematical iterative reconstruction method is used instead of the FDK algorithm. In other words, the sinogram of the decomposition material through the dual energy technique analysis generates various noise according to the analysis method, and utilizes a mathematical repetitive reconstruction method to remove the noise and restore the accurate image.

To do this, a mathematical iterative reconstruction method uses the sineogram's value matrix b and the matrix A related to the geometry of the system used to obtain the data to optimize and repeat the image x, . That is, as shown in the following equation (1), in the optimization process, the image is restored by minimizing the total variation (TV) of the estimated value x every time, and the degree of minimizing the total variation is called the parameter lambda (λ) to obtain a function f (x) for performing the iterative process.

Thereafter, the obtained function f (x) is solved through the mathematical repetition of Equation (2) to estimate an optimal image reconstruction value x. At this time, it is preferable to use a GP-BB (Gradient-Projection-Barzilai-Borwein) algorithm.

과 초기 조건을 설정한다. 이후, step 3에서는 반복 과정을 통해 추정된 값 x를 통해 반복 재구성의 스텝 사이즈를 결정하며, 이를 적용하여 step 4에서 수학적 반복 과정을 수행하여 최적화된 값 x를 추정한다.The mathematical repetition process will be described in more detail. In steps 1 and 2,

And initial conditions. In step 3, the step size of the repetitive reconstruction is determined through the estimated value x through the iterative process. Then, the mathematical iterative process is performed in step 4 to estimate the optimized value x.

The Cone beam artifact correction unit 200 then performs a forward projection operation for converting the Bone reconstruction image 210 of high density material estimated through the repeated reconstruction image reconstruction to projection data (Raw Data) Projection) and reconstructs it with the FDK algorithm to obtain the FDK bone reconstruction image 230 in which the Cone beam artifact appears.

Then, the reproduced cone beam artifact reconstructed image 230 is obtained through the difference between the bone reconstructed image 210 and the FDK bone reconstructed image 220.

Then, the cone beam artifact correction image 400 is finally obtained through a combination of the cone artifact generation image 300 and the cone beam artifact reproduction image 230.

Through this, we replace the threshold process of the existing two-pass algorithm by using a material distribution algorithm and an iterative reconstruction image reconstruction method that use two different energy spectra to estimate accurate high-density material.

As described above, according to the present invention, a high-density material image estimated through a dual-energy-based material decomposition algorithm and a mathematical repetitive reconstruction is performed in the same manner as the existing two-pass algorithm process to obtain a final concave artifact correction image 400 Through this method, it is unnecessary to designate an arbitrary threshold value as compared with the segmentation process by comparing the existing numerical values. Since the estimation performance of the high density material is constant and the iterative reconstruction method is used, it is resistant to noise .

Example

In order to implement the present invention by simulation, a cone beam CT system geometry is constructed as shown in FIG. At this time, MATLAB R2013b was used as the simulator.

In order to generate cone beam artifacts that occur according to the degree of offset in the z-axis direction from the center of the object 500, a large cylinder having X- Defrise phantom with 5 cylinders inserted with X-ray characteristic was set as simulation object (phantom).

The simulation parameters include a distance SO [mm] to the Iso-center of the X-ray source and the object 500, a distance OD from the Iso-center of the object 500 to the detector 600 [mm], the radius of the water cylinder R_Max [mm], the height of the water cylinder Z [mm], the radius of the bone cylinder R_Min [mm], the height of the bone cylinder A [mm] The number N of detectors 600, the size d [mm] of the detector 600, the image acquisition pixel size t [mm], the object 500 bone angle θ, and the detector 600 angle π.

In determining the simulation parameters, object truncation and maximum bone angle should be considered. Object truncation is a phenomenon that occurs when the object 500 is out of the area that can be covered by the detector 600. In this case, a truncation artifact occurs at the edge portion of the acquired CT image, . At this time, the condition that the object truncation does not occur is that the detector bone angle? Should be larger than the object bone angle?.

For the maximum detector bone angle, it is set to 10 degrees or more intentionally for comparison with the application of the cone beam CT system having a large area X-ray detector and other cone beam artifact reduction algorithms. Actually, the object cone angle θ = 10.12 °, The detector angle was set to π = 17.7 °.

The final system parameters used in the simulation are the distance between the X-ray source and the Iso-center of the object SO = 200 [mm], the distance from the Iso-center of the object to the detector OD = 200 [mm], the radius of the water cylinder R_Max = The height B of the bone cylinder A = 4 [mm], the distance B of the bone cylinder B = 3 [mm], the height of the water cylinder Z = 50 [mm], the radius of the bone cylinder R_Min = 40 [mm] The detector size d = 2 [mm], the image acquisition pixel size t = 1 [mm], the object bone angle [theta] = 10.12 [deg.], And the detector bone angle [pi] = 17.7 [deg.].

In addition, the damping coefficient (AC) of the material according to the dual energy spectrum (Low / High) used in the simulation and the smoothing parameter λ, nlet and Max_iteration No. used in the repetitive reconstruction restoration algorithm Table 1 shows the results. At this time, the nlet is one of the parameters of the FDK method, which is a conventional image restoration method. By setting a virtual let which acts as a detector between the detectors during image restoration, sufficient samples are acquired, The variable in the restoration algorithm that makes it possible to acquire.

In FIG. 6, different density is multiplied by a material density image of a defree phantom set for the purpose of comparison analysis between the existing two-pass algorithm and the algorithm proposed in the present invention The values of the damping coefficient of the bone cylinder were alternately set to be different. The defree phantom refers to a phantom most commonly used for generating cone beam artifacts as shown in FIG. 13, and a large cylindrical phantom (X- ray attenuation coefficient) of a plurality of small cylinders (substances having an X-ray attenuation coefficient equal to the bone in the case of the present patent) of another substance are spaced apart at regular intervals.

The reference phantom used in the simulation was constructed according to the system structure of FIG. 5, and each XZ-axis central profile 701 and the XY-axis central profile 702 are shown, as shown in FIG.

The reference phantom images 701 and 702 configured as described above are converted into projection data 703 that can be obtained through CT shooting through a forward projection process and the converted projection data 703 is converted to projection data 703 using the existing FDK algorithm (Backward Projection) process, and the image (704) in which the cone beam artifact appears in the reference phantom was obtained.

FIGS. 8 to 10 show simulations for a two-pass algorithm using existing numerical comparison. At this time, a threshold value for estimating a high-density material image was determined by analyzing the image value of the Cone beam artifact. In order to compare the performance of the two-pass algorithm according to the threshold value, The simulation was performed using the threshold value (0.5 / 1.0 / 1.5).

Based on the respective threshold values, as shown in FIGS. 8 to 10, a high-density material Segment ED Material (Bone) 706 is separated, and the estimated bone image is divided into a forward projection and a backward projection And acquires a cone beam artifact reconstructed image (Error Estimation) 707 through a difference between a bone segmentation image and a bone image in which a cone beam artifact appears.

Then, a cone beam artifact corrected image (Result) 708 is finally obtained through a combination of the scan data reconfiguration with FDK 705 and the Cone beam artifact reconstructed image 707.

As shown in FIGS. 8 to 10, in the conventional two pass algorithm, it can be seen that the variation of the degree of high density material image separated according to the threshold value is extremely large. Particularly, when the threshold value has a high threshold value (Threshold = 1.5), it can be seen that only two of the five bone cylinders manufactured by the estimated bone cylinder exhibit a bad partitioning performance so as to be estimated as a high-density material (Segment ED Material (Bone)) 706).

In particular, when the attenuation coefficient values of the bone cylinder are sequentially changed as in the case of the phantom set in advance, there are many cases in which the threshold value is very ambiguous. Even if the threshold value is set to a low threshold value, Due to the influence of the cone beam artifacts, the image is distorted and it is difficult to accurately estimate the high density material (refer to the Segment ED Material (Bone)) (706) Red line).

As such, there is a limit to explicit reproduction of the cone beam artifacts through the high-density material that is not accurately estimated, which affects the degree of cone beam artifact correction in the final image (Result) 708. [

On the other hand, in the algorithm proposed by the present invention, as shown in FIG. 11, a material decomposition algorithm using dual energy and a segmented ED material 802 are provided through a repetitive reconstruction image restoration process .

And the perfectly estimated bone image 802 is much more accurate in the Error Estimation 803 of the Cone Beam artifact and is ultimately more accurate than the conventional two- (804). ≪ / RTI >

12 is a diagram comparing the results of the present invention with the conventional two-pass algorithm.

The original FDK of Fig. 12 (a) is compared with the cone beam artifact correction image of Fig. 12 (b) obtained through the existing two-pass algorithm, Finally, it can be confirmed that the cone beam artifact correction image is excellent in cone beam artifact correction.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. It will be apparent to those skilled in the art that various modifications may be made without departing from the scope of the present invention. Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

Claims (12)

A dual energy technology analyzing unit for separating projection data obtained by using two different energies into projection data of only each constituent material including a Bone SaNogram and a Water Cygnogram through a material decomposition algorithm;
A forward projection that estimates a bone reconstruction image through image restoration using a mathematical repetitive reconstruction method using the bone tomogram and transforms the estimated bone reconstruction image into projection data (Raw Data) (FDK) reconstructed image obtained by reconstructing the FDK-BONE reconstructed image using the FDK-Boundary reconstruction image and a FDK-Boundary reconstruction image obtained by the FDK- And a cone beam artifact correction unit that finally obtains a cone beam artifact correction image through a combination of the cone beam artifact generation image and the cone beam artifact reconstruction image. The method according to claim 1,
Wherein the two different energies are two energy spectra and include a low energy spectrum and a high energy spectrum. 3. The method of claim 2,
The material decomposition algorithm constructs a polynomial through the projection data Y1 and Y2 already known through the dual energy spectrum, and outputs the corrected phantom estimating a coefficient of a polynomial through a calibration phantom and a least square fitting process. The method according to claim 1,
The mathematical iterative reconstruction method uses the sineogram value matrix b and the matrix A related to the geometry of the system used to acquire the data to calculate the total variation of x, ) Is minimized by an optimization process and an iterative process. 5. The method of claim 4, wherein the mathematical repetitive reconstruction method comprises:
In the optimization process, the image is restored by minimizing the total variation (TV) of the estimated value x every time, and the degree of minimizing the total variation is adjusted to the parameter lambda (λ) (x), and estimating an optimal image reconstruction value x through a mathematical repetition process using a GP-BB (Gradient-Projection-Barzilai-Borwein) algorithm. Wherein the cone beam artifact correction device comprises: (A) separating the projection data obtained by using the two energy sources through the dual energy technology analysis unit into projection data of only the constituent materials including the Bone-sirenogram and the water-cy Wow,
(B) estimating a bone between the projection data of the separated high density material (bone) through a cone beam artifact correction unit as a bone reconstruction image through image restoration using a mathematical iterative reconstruction method,
(C) reconstructing and correcting cone beam artifacts based on the bone reconstruction image of the high-density material estimated through the repeated reconstruction reconstruction of the cone beam artifact in the CT image. The method according to claim 6,
Wherein the two different energies are two energy spectra and include a low energy spectrum and a high energy spectrum. 7. The method of claim 6,
Constructing a polynomial through the projection data Y1 and Y2 already known through the dual energy spectrum,
Estimating a polynomial coefficient through a calibration phantom prepared in a state in which a value of a basis function is already known, and estimating the coefficient through a least square fitting process. Cone beam artifact correction method. 7. The method of claim 6, wherein the mathematically repetitive reconstruction method comprises:
An optimization process for minimizing the total variation (TV) value x of the image to be finally reconstructed using the value matrix b of the sinogram and the matrix A related to the geometry of the system used to acquire the data And estimating the converbean artifact in the CT image. 10. The method of claim 9, wherein the mathematical repetitive reconstruction method comprises:
In the optimization process, the image is restored by minimizing the total variation (TV) of the estimated value x every time, and the degree of minimizing the total variation is adjusted to the parameter lambda (λ) a first step of obtaining f (x)
And a second step of estimating an optimal image reconstruction value x through a mathematical repetition process using the GP-BB (Gradient-Projection-Barzilai-Borwein) algorithm. A method for correcting cone beam artifacts in an image processing apparatus. 11. The method of claim 10, wherein the second step
During the n-th iteration,

And setting an initial condition,
Determining a step size of an iterative reconstruction through an estimated value x through the iterative process;
And estimating an optimized value x by performing the iterative process. ≪ RTI ID = 0.0 > [10] < / RTI > 7. The method of claim 6, wherein step (C)
A forward projection process for transforming a bone reconstruction image of a high density material estimated by the repeated reconstruction image reconstruction into projection data (Raw Data) obtained through CT imaging, and restoring the image by the FDK algorithm, Acquiring an FDK bone reconstruction image,
Acquiring a cone beam artifact reconstructed image reproduced through a difference between the bone reconstructed image and the FDK bone reconstructed image;
And finally obtaining a cone beam artifact corrected image through a combination of the cone beam artifact occurrence image and the obtained cone beam artifact reproduced image.

Images